In my efforts to make it possible to resume a failed clone, I discovered
something quite  interesting. When I  first started the work,  I noticed
that clones  of my local Fossil  repository were taking a  long time and
thought it  was some inefficiency  that I had  added with my  changes. I
spent quite a few hours trying to debug my code and then I realized that
it wasn't my code at all. It was the fact that my repository was old and
that I hadn't rebuilt  it in a long time and so it  didn't have a lot of
delta compression as a result.

My Fossil repository was close to 350MB and so it would take a long time
and many round-trips of clone protocol  before it would complete. Then I
realized that the reason why my repository was so big---and consequently
took so long to clone---was because over the years it had received a lot
of uncompressed  content. So  I ran "fossil  rebuild --compress"  and it
reduced the  repository size significantly  (down to 64MB  approx.), and
consequently  my clones  happen  much  more quickly  now.  I suppose  in
retrospect I could have just  run "fossil rebuild --compress-only" but I
was in a hurry and didn't notice that option.

Long story short,  for those who publish Fossil repositories  with a lot
of commits, you can improve the user experience for those who clone your
repository by running "fossil rebuild --compress-only" every so often.

Andy